Reference frames in virtual spatial navigation are viewpoint dependent

Abstract

Spatial navigation in the mammalian brain relies on a cognitive map of the environment. Such cognitive maps enable us, for example, to take the optimal route from a given location to a known target. The formation of these maps is naturally influenced by our perception of the environment, meaning it is dependent on factors such as our viewpoint and choice of reference frame. Yet, it is unknown how these factors influence the construction of cognitive maps. Here, we evaluated how various combinations of viewpoints and reference frames affect subjects' performance when they navigated in a bounded virtual environment without landmarks. We measured both their path length and time efficiency and found that (1) ground perspective was associated with egocentric frame of reference, (2) aerial perspective was associated with allocentric frame of reference, (3) there was no appreciable performance difference between first and third person egocentric viewing positions and (4) while none of these effects were dependent on gender, males tended to perform better in general. Our study provides evidence that there are inherent associations between visual perspectives and cognitive reference frames. This result has implications about the mechanisms of path integration in the human brain and may also inspire designs of virtual reality applications. Lastly, we demonstrated the effective use of a tablet PC and spatial navigation tasks for studying spatial and cognitive aspects of human memory.

Orientation and metacognition in virtual space.

“…As outlined above, previous research consistently suggests a profound and lasting effect of the experienced perspective (e.g., Kelly & McNamara, 2010;Montello et al, 2004;Shelton & McNamara, 2004;Török et al, 2014). The direct navigation experience in the tunnel task, i.e., the visual perception of travelling through a tunnel (rather than viewing it from above), should therefore have a strong impact on the cognitive representation, overruling other possible reference systems that do not correspond to the experienced view.…”

“…Correspondingly, different perspectives shown in a virtual environment lead to different kinds of cognitive representations (Shelton & McNamara, 2004;Török et al, 2014). With an egocentric perspective on the scene, it is easier to navigate; with an aerial view, path integration relative to environmental cues is easier.…”

“…Problems arise in particular when the view on the scene is ground-based rather than aerial, yet not egocentric, as in a third-person perspective using an avatar with a constantly north-oriented map. The mismatch of perceived view and cognitive reference system then affects both navigation and orientation (Török et al, 2014).…”

“…Associating these concepts with each other is actually a hypothesis about cognitive mechanisms. In particular, it is hypothesised that the cognitive strategies activated by the perceptually sparse tunnel task are, at least to some extent, the same as those activated by tasks that involve views of decisions in the real world triggered by environmental cues (Kelly et al, 2008;Török et al, 2014), verbalisations of internalised environmental knowledge, embodied experience, and so on. However, this fairly far-reaching hypothesis has so far not been addressed explicitly.…”

“…Studying navigation in virtual reality (VR) is becoming increasingly popular in spatial cognition (Boccia, Nemmi, & Guariglia, 2014;Maguire et al, 1999, Darken & Peterson, 2002, Stankiewicz et al, 2006, Török et al, 2014, even for investigating primates (Dolins et al, 2014) or rodents (Harvey et al, 2009;Chen et al, 2013). There are obvious advantages; for instance, features of the environment (e.g.…”